Abstract
We will focus on a problem of estimating an unknown regression function based on spline approach, while taking into account also some sudden jumps - change-points. We will discuss some basic methods and their properties for a simple spline approach and afterwards we will extend them and we will implement an algorithm which would allow for change- points occurrences at some certain, pre-defined points - so called knots.
The proposed methods are applied to B19 parvovirus data in Belgium population and we have also used a simulation study for additional conclusions.